OmniSVG: Unified model for high-quality SVG generation using pre-trained Vision-Language Models. Overcomes limitations of existing methods by producing complex SVGs efficiently. Introduces MMSVG-2M dataset with 2M annotated SVGs for enhanced generation. Supports multiple modes like Text-to-SVG and Image-to-SVG. Outperforms current approaches, showing promise in professional design use.
OmniSVG: a Unified Scalable Vector Graphics Generation Model
